NetBooster: Empowering Tiny Deep Learning By Standing on the Shoulders of Deep Giants
DescriptionTiny deep learning has attracted tremendously increased attention driven by the substantial demand for deploying deep learning on numerous intelligent Internet-of-Things (IoT) devices. However, it is still challenging to unleash tiny deep learning's full potential on both large-scale datasets and downstream tasks due to the under-fitting issue caused by the limited model capacity of tiny neural networks (TNNs). To this end, we propose a framework called NetBooster to empower tiny deep learning by augmenting the architecture of TNNs via an expansion-then-contraction training strategy. Extensive experiments show that our NetBooster consistently outperforms state-of-the-art tiny deep learning solutions.
Event Type
Research Manuscript
TimeWednesday, July 12th10:40am - 10:55am PDT
Location3004, 3rd Floor
AI/ML Algorithms